ErrorMap and ErrorAtlas: Charting the Failure Landscape of Large Language Models

TL;DR

ErrorMap and ErrorAtlas provide a novel framework for understanding the underlying causes of failures in large language models, enabling more targeted debugging and evaluation beyond traditional success metrics.

Contribution

We introduce ErrorMap, a method to identify and analyze the sources of LLM failures, and ErrorAtlas, a comprehensive taxonomy of error types across multiple datasets and models.

Findings

ErrorAtlas reveals common failure patterns like omission and misinterpretation.

ErrorMap can be applied to any model or dataset for failure analysis.

The approach uncovers underexplored error types in LLM research.

Abstract

Large Language Models (LLM) benchmarks tell us when models fail, but not why they fail. A wrong answer on a reasoning dataset may stem from formatting issues, calculation errors, or dataset noise rather than weak reasoning. Without disentangling such causes, benchmarks remain incomplete and cannot reliably guide model improvement. We introduce ErrorMap, the first method to chart the sources of LLM failure. It extracts a model's unique "failure signature", clarifies what benchmarks measure, and broadens error identification to reduce blind spots. This helps developers debug models, aligns benchmark goals with outcomes, and supports informed model selection. ErrorMap works on any model or dataset with the same logic. Applying our method to 35 datasets and 83 models we generate ErrorAtlas, a taxonomy of model errors, revealing recurring failure patterns. ErrorAtlas highlights error types that are currently underexplored in LLM research, such as omissions of required details in the output and question misinterpretation. By shifting focus from where models succeed to why they fail, ErrorMap and ErrorAtlas enable advanced evaluation - one that exposes hidden weaknesses and directs progress. Unlike success, typically measured by task-level metrics, our approach introduces a deeper evaluation layer that can be applied globally across models and tasks, offering richer insights into model behavior and limitations. We make the taxonomy and code publicly available with plans to periodically update ErrorAtlas as new benchmarks and models emerge.